i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/he

Accurate experimental (p, r, T) data of natural gas

mixtures for the assessment of reference equations
of state when dealing with hydrogen-enriched
natural gas

Roberto Herna
ndez-Go
mez a, Dirk Tuma b, Daniel Lozano-Martı́n a,

C
esar R. Chamorro a,*
a
Grupo de Termodina
mica y Calibraci
on (TERMOCAL), Dpto. Ingenierı́a Energ
etica y Fluidomeca
nica, Escuela de
Ingenierı́as Industriales, Universidad de Valladolid, Paseo del Cauce, 59, E-47011 Valladolid, Spain
b
BAM Bundesanstalt für Materialforschung und -prüfung, D-12200 Berlin, Germany

article info abstract

Article history: The GERG-2008 equation of state is the approved ISO standard (ISO 20765-2) for the
Received 31 July 2018 calculation of thermophysical properties of natural gas mixtures. The composition of
Received in revised form natural gas can vary considerably due to the diversity of origin. Further diversification was
3 October 2018 generated by adding hydrogen, biogas, or other non-conventional energy gases. In this
Accepted 5 October 2018 work, high-precision experimental (p, r, T) data for two gravimetrically prepared synthetic
Available online 2 November 2018 natural gas mixtures are reported. One mixture resembled a conventional natural gas of 11
components (11 M) with a nominal mixture composition (amount-of-substance fraction) of
Keywords: 0.8845 for methane as the matrix compound with the other compounds being 0.005 for
Hydrogen-enriched natural gas oxygen, 0.04 for nitrogen, 0.015 for carbon dioxide, 0.04 for ethane, 0.01 for propane, 0.002
Gravimetric preparation each for n- and isobutane, and ultimately 0.0005 each for isopentane, n-pentane, and n-
Single-sinker densimeter hexane. The other mixture was a 13-component hydrogen-enriched natural gas with a low
Reference equations of state calorific value featuring a nominal composition of 0.7885 for methane, 0.03 for hydrogen,
GERG-2008 0.005 for helium, 0.12 for nitrogen, 0.04 for carbon dioxide, 0.0075 for ethane, 0.003 for
AGA8-DC92 propane, 0.002 each for n- and isobutane, and 0.0005 each for neopentane, isopentane, n-
pentane, and n-hexane. Density measurements were performed in an isothermal opera-
tional mode at temperatures between 260 and 350 K and at pressures up to 20 MPa by using
a single-sinker densimeter with magnetic suspension coupling. The data were compared
with the corresponding densities calculated from both GERG-2008 and AGA8-DC92 equa-
tions of state to test their performance on real mixtures. The average absolute deviation
from GERG-2008 (AGA8-DC92) is 0.027% (0.078%) for 11 M and 0.095% (0.062%) for the 13-
component H2-enriched mixture, respectively. The corresponding maximum relative de-
viation from GERG-2008 (AGA8-DC92) amounts to 0.095% (0.127%) for 11 M and 0.291%
(0.193%) for the H2-enriched mixture.
© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

* Corresponding author.
E-mail address: cescha@eii.uva.es (C.R. Chamorro).
https://doi.org/10.1016/j.ijhydene.2018.10.027
0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
21984 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8
Introduction
The increasing global energy demand with its impact on
mankind is a growing imperative for the development of
alternative energy sources which truly fulfil the condition of
sustainability. Initiated by the first oil crisis in 1973 [1],
hydrogen is promoted as a sustainable alternative energy
carrier in the implementation of energy policies that do not
depend on fossil resources, or, in short, on carbon. The last
years have seen a vibrant and continuously growing activity
on all fields related. The concept of “decarbonisation” literally
means to replace carbon atoms with hydrogen atoms. The
book entitled “The Hydrogen Economy: Opportunities and
Challenges” by Ball and Wietschel published in 2009 high-
lights hydrogen comprehensively from a technical, environ-
mental, and socioeconomic perspective [2]. At present, the
global economy still runs largely on fossil fuels. In a long-term
perspective, fossil fuels must develop away from energy into
the raw material for producing chemical products. Hydrogen,
however, naturally occurs in a bonded form and thus has first
to be released by using energy, ideally from renewable sources
[3e7]. An illustrative example is the so-called “power-to-gas”
technique, that is the use of surplus electrical power to pro-
duce a gas fuel, where hydrogen is made by electrolysis [8e10].
One hydrogen-related approach that has been put forward for
making a transition towards a carbon-free economy is adding
hydrogen into existing natural gas transport and distribution
systems [11e15]. Capocelli and De Falco have put it in an
introductory book chapter “Enriched methane is a ready so-
lution for the transition towards the hydrogen economy” [16];
in other words, enriched methane is a pathway to introduce
hydrogen in our established and consolidated energy
infrastructure.
Apart from plain economic figures, various technical as-
pects over the entire chain, such as safety, availability, stor-
age, metering, equipment and operational parameters,
efficiency, as well as material characteristics, have to be suc-
cessfully addressed towards a full affirmation. These research
activities manifested in a notable number of studies. Some
recently published papers which particularly address the
aforementioned issues for enriched methane are: experi-
ments of turbulent explosions, focusing on the influence on
mixture reactivity by Li et al. [17] and explosion studies of
methaneehydrogen mixtures by Lowesmith et al. [18] on
safety; a study on transition pathways to future energy
infrastructure by Maroufmashat and Fowler on availability
[19]; development of measuring standards to support fiscal
metering by van der Veen et al. on analysis and metering [20];
a comprehensive review on combustion engines for H2-
enriched natural gas by Mehra et al. [21], a study of low-
calorific value coal gas combustion by Karyeyen and Ilbas _ ‚
[22,23], and the studies investigating the effects of composi-
tion on engine performance, combustion, and emission by
Hora and Agarwal [24] and Cellek and Pınarbas‚ı [25], respec-
tively, deal with equipment and operation.
Due to the low calorific value of hydrogen on a volumetric
basis, a virtual hydrogen-spiked energy gas would require a
gas with about 80% of hydrogen to reduce the emission of
carbon dioxide by 50% [11]. The approved ISO standard (ISO
20765-2) for the calculation of thermophysical properties of
established natural gas-type mixtures is the GERG-2008
equation of state [26,27]. The addition of hydrogen will
contribute further diversification to the composition of those
mixtures. The performance of equation-of-state (EoS) models,
such as the GERG-2008 EoS or the AGA8-DC92 EoS developed
by the American Gas Association [28], must be validated using
consolidated volumetric (pVT) data of high accuracy. Real
mixtures that are prepared by gravimetry according to ISO
6142-1 display the highest accuracy in composition and thus
qualify best [29].
Prior to this work, Richter et al. have investigated three H2-
enriched natural gases [30]. The mixtures for this study were
prepared from a 21-component high-calorific pipeline gas that
was blended with hydrogen to a hydrogen content of
approximately 5, 10, and 30 mol-%. Volumetric data were
recorded at temperatures between 273 and 293 K and up to a
maximum pressure of 8 MPa using a two-sinker densimeter.
The experimental density data were compared with calcu-
lated values from both the GERG-2008 and the AGA8-DC92
EoS. For both EoS, the relative deviations in density remained
below 0.05% for the 5- and 10-% mixture but raised to a value
of about 0.1% for the 30-% mixture. The authors explained
latter with larger errors in the composition analysis at a high
hydrogen content.
In another volumetric study on mixtures, Atilhan, Apar-
icio, Hall, and co-authors have employed a (H2-free) deep-
water natural gas mixture with heavy hydrocarbon content
(i.e., C6þ up to n-C9 > 0.2 mol-%) that were gravimetrically
prepared and traceable to the National Institute of Standards
and Technology NIST [31]. Since the authors intended to
explore the limits of the equation-of-state models (both GERG-
2008 and AGA8-DC92), the experiments were conducted at
temperatures between 270 and 340 K and up to pressures of
35 MPa using a magnetic suspension densimeter. They found
density deviations larger than 0.1% especially for pressures <
10 MPa in the vicinity of the phase envelope for both models.
Chapoy and co-workers determined densities and speed-
of-sound values of a synthetic natural gas (88 mol-% of
methane) with certified composition at temperatures between
323 and 415 K and pressures up to 58 MPa [32] using a high-
pressure and high-temperature vibrating-tube densimeter
and a specially adapted in-house made acoustic cell. Addi-
tionally, an isochoric cell was employed to cover the entire
operational range. In addition to GERG-2008 and AGA8-DC92
EoS, the (cubic) Peng-Robinson [33] and the Soave-Benedict-
Webb-Rubin [34] EoS were used to evaluate the experimental
data and to calculate various derivable thermodynamic
properties. The Peng-Robinson EoS featured a maximum de-
viation in the predicted density of about 2.8%, whereas the
maximum deviation for the other models did not exceed 0.7%.
This work provides new high-precision experimental (p, r,
T) data for two gravimetrically prepared natural gas mixtures.
The nominal composition of these mixtures is given in Table
1. Both mixtures also qualify as primary calibration stan-
dards. The first mixture resembled a conventional natural gas
of 11 components and is labeled BAM-G 420 or 11 M according
to the specification given in the directive PTB-A 7.63 by the
Physikalisch-Technische Bundesanstalt PTB [35]. The second
mixture was a 13-component H2-enriched natural gas mixture
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8 21985

with a low calorific value to facilitate support to power-to-gas
applications which was proposed by the Consultative Com-
mittee for Amount of Substance: Metrology in Chemistry and
Biology (CCQM) of the Bureau International des Poids et
Mesures (BIPM) for the interlaboratory key comparison K 118
[36]. Density measurements were performed at temperatures
between 260 and 350 K and at pressures up to 20 MPa using a
single-sinker densimeter with magnetic suspension coupling.
94 data points were recorded for 11 M and 99 for the H2-
enriched gas mixture, respectively, in an isothermal opera-
tional mode at 260, 275, 300, 325, and 350 K. The data were
compared with the corresponding densities calculated from
GERG-2008 and AGA8-DC92 EoS, via REFPROP [37], and can
thus give an assessment for the performance of these models
on novel compositions of energy gases.
Experimental
Gas mixture preparation
Both natural gas mixtures were prepared gravimetrically by
the Federal Institute for Materials Research and Testing
(Bundesanstalt für Materialforschung und -prüfung, BAM) in
Berlin, Germany, according to the procedures outlined in ISO
6142-1 [29].
The nominal composition of the two mixtures investigated
as well as purity, supplier, molar mass, and critical
parameters of the pure compounds are given in Table 1. Table
2 shows the cylinder identifiers, gravimetric composition, and
the corresponding (absolute) expanded uncertainty (k ¼ 2) of
the mixtures. The prepared mixtures were supplied in
aluminum cylinders of a volume of 10 dm3. All compounds
were used without further purification, but information on
impurities from the specification was considered in the
mixture preparation.
Due to their particular composition, the preparation of the
final mixtures required several premixtures and the transfer
of liquid and liquefied compounds via high-pressure mini-
cylinders. Liquid compounds of low volatility, such as n-hex-
ane, were directly introduced by gas-tight glass syringes. A
detailed chart for the individual preparation is given in the
Supporting Information. In our method, pressure difference is
the exclusive driving force for the mass transfer into the
recipient cylinder. Optionally, heating had to be administered
to create sufficient pressure differences as well as to avoid
condensation and the assigned portions of liquid and liquefied
compounds were always vaporized from the minicylinders
during the transfer. The complete transfer of vapors and liq-
uids from the tubing into the recipient cylinder required a
purge gas that in turn was a pure gas or a mixture. The use of
minicylinders for smaller mass portions ensured that the
corresponding uncertainty of the weighed mass differences
did not exceed the preset qualifying limits for the preparation
of reference materials. During the filling procedure, the
recipient cylinder freely stood on an electronic comparator

Table 1 e Nominal composition of the two gas mixtures, purity, supplier, molar mass, and critical parameters of the
individual components.
Components Synthetic natural gas H2-enriched natural Purity Supplier M/g$mol1 Critical
11 M (BAM-G 420) gas CCQM K 118 (mol-%) parametersa
xi (102 mol$ Uðxi Þ (k ¼ 2)b xi (102 mol$ Uðxi Þ (k ¼ 2)c Tc/K pc/MPa
1 2 1 1
mol ) (10 mol$mol ) mol ) (102 mol$mol1)
Methane 88.45 0.044 78.85 e 99.9995 Linded 16.043 190.56 4.60
Ethane 4.0 0.016 0.75 e 99.999 Mathesone 30.069 305.32 4.87
Propane 1.0 0.004 0.30 e 99.999 Scottf 44.096 369.89 4.25
Butane 0.2 0.001 0.20 e 99.95 Scottg 58.122 425.13 3.80
Isobutane 0.2 0.001 0.20 e 99.98 Scotth 58.122 407.81 3.63
Pentane 0.05 0.00025 0.05 e 99.7 Sigmai 72.149 469.70 3.37
Isopentane 0.05 0.00025 0.05 e 99.7 Sigmai 72.149 460.35 3.38
Neopentane e e 0.05 e 99.0 Linded 72.149 433.74 3.20
Hexane 0.05 0.00025 0.05 e 99.7 Sigmai 86.175 507.82 3.03
Carbon dioxide 1.5 0.008 4.0 e 99.9995 Air Liquidej 44.010 304.13 7.38
Nitrogen 4.0 0.012 12.0 e 99.9995 Linded 28.013 126.19 3.39
Oxygen 0.5 0.003 e e 99.9999 Westfalenk 31.999 154.58 5.04
Helium e e 0.50 e 99.9995 Linded 4.003 126.19 3.40
Hydrogen e e 3.0 e 99.9999 Linded 2.016 33.15 1.30
a
Critical parameters were obtained by using the default equation for each substance in REFPROP software [37].
b
The given uncertainty is certified by BAM on request according to the PTB document PTB-A 7.63 [35].
c
CCQM agreed on a nominal target composition only [36].
d
Linde AG, Unterschleißheim, Germany.
e
Matheson Tri-Gas, Inc., Montgomeryville PA, USA.
f
Scott Specialty Gases BV, Breda, The Netherlands.
g
Scott UK, Newcastle-under-Lyme, UK.
h
Scott Specialty Gases, Inc., Plumsteadville PA, USA.
i
Sigma-Aldrich Chemie GmbH, Steinheim, Germany.
j
Air Liquide AG, Düsseldorf, Germany.
k
Westfalen AG, Münster, Germany.
21986 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8

Table 2 e Gravimetric composition of the natural gas mixtures studied including information on impurities from the
supplier. Impurity compounds are marked in italic type.
Component 11 M (BAM-G 420) natural gas H2-enriched natural gas
BAM no.: C49358-090825 CCQM K 118
BAM no.: 8099-160905
xi /(102 mol$ Uðxi Þ(k ¼ 2)/ xi /(102 mol$ Uðxi Þ(k ¼ 2)/
mol1) (102 mol$mol1) mol1) (102 mol$mol1)
Methane 87.663627 0.003485 78.821237 0.003809
Ethane 4.225210 0.000462 0.757359 0.000158
Propane 1.049010 0.002110 0.297078 0.000089
Butane 0.212654 0.000101 0.200439 0.000098
Isobutane 0.210325 0.000084 0.197953 0.000035
Pentane 0.051829 0.000027 0.050134 0.000021
Isopentane 0.052184 0.000027 0.049928 0.000021
Neopentane e e 0.049615 0.000031
Hexane 0.052567 0.000024 0.050708 0.000019
Carbon dioxide 1.622854 0.000302 4.001075 0.000283
Nitrogen 4.321699 0.000783 12.017829 0.000767
Oxygen 0.538012 0.000107 e e
Helium e e 0.496897 0.000297
Hydrogen e e 3.009733 0.001323
Hydrogen 0.000003 0.000002 e e
Argon 0.000001 0.000001 e e
Ethene 0.000003 0.000002 0.000001 < 0.000001
Neopentane 0.000019 0.000004 e e
Carbon monoxide 0.000003 0.000003 0.000002 0.000002
Oxygen e e 0.000011 0.000006

balance (Sartorius LA 34000P-0CE, Sartorius AG, Go € ttingen,
Germany, weighing range: 34 kg, readability: 0,1 g) to simul-
taneously monitor the gas feed stream. The exact mass of the
gas added into the recipient cylinder was determined after
each filling step using a high-precision mechanical balance
(Voland model HCE 25, Voland Corp., New Rochelle NY, USA,
weighing range: 25 kg, readability: 2.5 mg). The small masses,
such as minicylinders and syringes, were determined on an
electronic comparator balance (Sartorius CCE 2004, weighing
range: 2500 g, readability:  0.2 mg). After the last filling step,
each mixture was always homogenized by subsequent heat-
ing and rolling the cylinder.
Prior to density determination at the University of Valla-
dolid, the two natural gas mixtures were validated at BAM by
gas chromatography (GC) on a multichannel (12 in total for 6
trains) process analyzer (Siemens MAXUM II, Siemens AG,
Karlsruhe, Germany). The equipment was specifically
designed and configured to analyze natural gas mixtures
which can in general contain methane (as main compound),
ethane, ethene, propane, propene, isobutane, n-butane, neo-
pentane, isopentane, n-pentane, n-hexane, nitrogen, oxygen,
carbon dioxide, carbon monoxide, helium, and hydrogen. The
configuration as a process GC required the individual con-
centrations of the analyzable compounds to be within a pre-
defined range. The GC operated in a single isothermal mode at
60  C. Each channel was equipped with packed columns and a
separate thermal conductivity detector (TCD). Helium and
hydrogen were analyzed on a single channel operated by ni-
trogen as the carrier gas. The other channels were operated by
helium and attributed to the analysis of a cluster of com-
pounds. (There were also “non-analyzing” separate channels
which served several backflush procedures and guard col-
umns, respectively.) The compound clusters that shared one
channel were combined as follows (with ascending retention
times): methane, carbon dioxide, ethene, ethane d propane,
propene d oxygen, nitrogen, carbon monoxide d isobutane,
n-butane, neopentane, isopentane, n-pentane d n-hexane.
The validation was performed according to ISO 12963,
preferably by a two-point bracketing method using indepen-
dently prepared calibration gases [38]. The 11 M mixture was
frequently used as calibration gas (BAM-G 420) in routine
certifications, the H2-enriched natural gas mixture was vali-
dated during the validation campaign for the samples to be
distributed to the participants in the interlaboratory key
comparison K 118 in which BAM acted as pilot lab. The results
of the GC analysis and the corresponding (gravimetric)
composition of the appropriate gas mixtures used for valida-
tion go with Table 3. There, a mixture regularly employed in
certification analyses by a direct match method is given for
the 11 M mixture and both mixtures to perform a two-point
bracketing method for the H2-enriched natural gas mixture,
respectively. The deviations between gravimetric composition
and that from GC analysis were sufficiently low to pass the
criteria for certification.
Apparatus and method
The volumetric (p, r, T) data were recorded with a single-
sinker densimeter with magnetic suspension coupling. That
particular method operates on the Archimedes’ principle. A
magnetic suspension coupling system enables to determine
the buoyancy force on a sinker immersed in the medium so
that density values over a large range of temperature and
pressure can be measured with high accuracy [39]. The design
had been adapted and optimized for density measurements of
both pure gases and gaseous mixtures by researchers from the
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8 21987

Table 3 e Results of the GC analysis and relative deviation between gravimetric composition and GC analysis for the two
natural gas mixtures. The results are followed by the gravimetric composition of the employed validation mixtures
(arranged in the same way as in Table 2).
11 M (BAM-G 420) natural gas H2-enriched natural gas
Component BAM no.: C49358-090825 CCQM K 118
BAM no.: 8099-160905
Concentration Relative deviation Concentration Relative deviation
between between
gravimetric gravimetric
preparation preparation
and GC analysis and GC analysis
xi /(102 mol$ Uðxi Þ(k ¼ 2)/ % xi /(102 mol$ Uðxi Þ(k ¼ 2)/ %
mol1) (102 mol$mol1) mol1) (102 mol$mol1)
Methane 87.6348 0.0450 0.033 78.8143 0.0548 0.009
Ethane 4.2187 0.0086 0.155 0.7571 0.0046 0.036
Propane 1.0478 0.0022 0.113 0.2969 0.0009 0.045
Butane 0.2124 0.0002 0.131 0.2008 0.0002 0.172
Isobutane 0.2100 0.0002 0.135 0.1974 0.0002 0.292
Pentane 0.0518 0.0002 0.031 0.0501 0.0002 0.024
Isopentane 0.0522 0.0001 0.049 0.0499 0.0001 0.038
Neopentane e e e 0.0495 0.0001 0.146
Hexane 0.0525 0.0003 0.106 0.0507 0.0001 0.048
Carbon dioxide 1.6242 0.0057 0.083 4.0005 0.0091 0.014
Nitrogen 4.3130 0.0017 0.200 12.0168 0.0037 0.008
Oxygen 0.5377 0.0005 0.062 e e e
Helium e e e 0.4969 0.0007 0.005
Hydrogen e e e 3.0105 0.0030 0.025
Validation mixture Validation mixture
BAM no.: 96054 925-090811 BAM no.: 8056-140922
Methane 89.021569 0.003176 80.443530 0.003245
Ethane 3.807373 0.000420 0.700350 0.000141
Propane 0.945272 0.001901 0.274717 0.000080
Butane 0.190906 0.000092 0.184936 0.000091
Isobutane 0.190061 0.000077 0.186453 0.000035
Pentane 0.049116 0.000025 0.046127 0.000055
Isopentane 0.048774 0.000025 0.046163 0.000057
Neopentane e e 0.045306 0.000057
Hexane 0.047765 0.000022 0.046238 0.000083
Carbon dioxide 1.426743 0.000279 3.700223 0.000247
Nitrogen 3.799404 0.000718 11.103398 0.000655
Oxygen 0.472990 0.000097 e e
Helium e e 0.471769 0.000280
Hydrogen e e 2.750777 0.001192
Hydrogen 0.000003 0.000002 e e
Argon 0.000001 0.000001 e e
Ethene 0.000003 0.000002 0.000001 < 0.000001
Neopentane 0.000018 0.000004 e e
Carbon monoxide 0.000003 0.000003 0.000002 0.000002
Oxygen e e 0.000011 0.000006
Validation mixture
BAM no.: 8045-140923
Methane 77.208758 0.004338
Ethane 0.810092 0.000181
Propane 0.321367 0.000100
Butane 0.216555 0.000107
Isobutane 0.215044 0.000041
Pentane 0.054141 0.000064
Isopentane 0.054195 0.000067
Neopentane 0.052772 0.000067
Hexane 0.053205 0.000096
Carbon dioxide 4.304190 0.000318
Nitrogen 12.928316 0.000872
Oxygen e e
Helium 0.535826 0.000322
Hydrogen 3.245527 0.001441
(continued on next page)
21988 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8
Table 3 e (continued )
11 M (BAM-G 420) natural gas
Component BAM no.: C49358-090825
Concentration
xi /(102 mol$ Uðxi Þ(k ¼ 2)/
mol1) (102 mol$mol1)
Hydrogen
Argon
Ethene
Neopentane
Carbon monoxide
Oxygen
University of Bochum, Germany [40,41]. For the specific details
of the apparatus used in this work the reader is referred to the
paper by Monde
jar et al. and the references cited therein [42].
The sinker used here had a cylindrical shape and was made
of silicon with a mass of 61.59181 ± 0.00016 g and a corre-
sponding volume of 26.444 ± 0.003 cm3 (k ¼ 2), determined at
T ¼ 293.05 K and p ¼ 1.01134 bar. The device operated with a
load compensation system that consisted of two calibrated
masses (provided by Rubotherm GmbH, Bochum, Germany)
made of tantalum and titanium, respectively, that have
approximately the same volume (4.9 cm3) but different masses.
Mass and volume of these mass pieces underwent a calibration
at the Mass Division of the Spanish National Metrology Insti-
tute (Centro Espan ~ ol de Metrologı́a, CEM) prior to their use [43].
The weight difference of both masses resembles that of the
sinker. The characteristic load compensation allows for
running the measurements near the zero point of the balance
where the effect of air buoyancy becomes negligible. A detailed
description of operational procedures and data reduction to
obtain the final result for the density value is given in the pa-
pers by Monde
jar et al. [42] for our equipment and by Richter
et al. [44] and McLinden [45] on general aspects. Principally,
density r is calculated using Eq. (1)


mS0  mSf
r¼
VS ðT; pÞ
where the numerator represents the buoyancy force that is
exerted on the sinker. The mass mS0 stands for the weighing
result of the sinker in vacuum, mSf for the corresponding
result in a pressurized medium. The weighing operation is
done with a high-precision electronic microbalance (Mettler
Toledo XPE205DR, Mettler Toledo GmbH, Gießen, Germany,
normal weighing range: 81 g, readability: 0.01 mg, extended
weighing range: 220 g, readability: 0.1 mg). The denominator
VS (T, p) of Eq. (1) is the pressure- and temperature-dependent
volume of the sinker immersed in the respective medium. The
temperature in the cell is determined by two platinum resis-
tance thermometers (S1059PJ5X6, Minco Products, Inc., Min-
neapolis MN, USA) and a reference resistance which are
connected with an AC comparator resistance bridge (F700,
H2-enriched natural gas
CCQM K 118
BAM no.: 8099-160905
Relative deviation Concentration Relative deviation
between between
gravimetric gravimetric
preparation preparation
and GC analysis and GC analysis
% xi /(102 mol$ Uðxi Þ(k ¼ 2)/ %
mol1) (102 mol$mol1)
e e
e e
0.000001 < 0.000001
e e
0.000002 0.000002
0.000011 0.000005
Automatic Systems Laboratories, Redhill, England). The
pressure is recorded by two pressure transducers for the
entire operational range, namely a Paroscientific 2500A-101
for pressures from 0 to 3 MPa and a Paroscientific 43 KR-HHT-
101 (Paroscientific Inc., Redmond WA, USA) for elevated
pressures up to 20 MPa. To achieve a higher accuracy, evalu-
ation of data from the single-sinker technique requires a
correction for two effects that occur, i.e., the force trans-
mission error (FTE) due to magnetic coupling [46] and ab-
sorption of gas molecules on the surface inside the cell and
sinker [47]. The FTE is a combination of two effects, namely an
apparatus effect and a medium-specific effect. The apparatus
effect can be eliminated by simply determining the sinker
mass in vacuum mS0 after finishing an isotherm (see also
Section Experimental). However, the medium-specific effect
results from the magnetic susceptibilities (c) of the mixtures.
Literature recommends a consideration only for strongly
paramagnetic fluids [46], such as oxygen or oxygen mixtures.
The susceptibility value of the studied gas mixtures,
composed mainly of methane, estimated with the additive
law proposed by Bitter [48], is three orders of magnitude lower
than the magnetic susceptibility of oxygen (cO2 ¼ 1.78$106).
Additionally, the sinker density also contributes to the
medium-specific effect, with low-density sinkers being pref-
erable [46]. In our work, the silicon sinker is of low density
(1)
(rSi ¼ 2329 kg m3) compared to the titanium sinkers
(rTi ¼ 4506 kg m3) used by other single-sinker densimeters.
The low susceptibility values of the studied gas mixtures in
conjunction with the low density of the silicon sinker make a
medium-specific correction negligible that was therefore dis-
carded. The influence of sorption effects inside the measuring
cell is less defined and thus more difficult to specify due to
particular interactions between the medium and the surface.
Other researchers working with similar techniques reported
errors in density up to 0.1% from such phenomena [47]. The
best and approved way to minimize that effect is a procedure
of several alternating flushing and evacuating the measuring
cell before the actual measurement is started. The residence
time inside the cell of the mixture to be investigated did never
exceed a period of 40 h. Another viable way to obtain more
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8 21989

40 40
GERG-2008 EoS GERG-2008 EoS

30 30

p/MPa
p/MPa
20 20
Industry Industry
10 10

0 0
0 100 200 300 400 500 0 100 200 300 400 500
a) T/K
b) T/K

Fig. 1 e p, T-phase diagram showing the experimental points measured () and the calculated saturation curve for the a)
11 M natural gas-like mixture and b) H2-enriched natural gas mixture. The marked temperature and pressure ranges
represent the validity of the GERG-2008 EoS (dashed line) and the area of interest for the transport and compression of gas
fuels (thin dashed line).

information on those sorption effects was a specific sorption
test for the particular gas mixture in the same way as
executed in previous studies [49,50].
Experimental
The volumetric (p, r, T) data were recorded as isotherms. Both
mixtures were investigated at temperatures of 260, 275, 300,
325, and 350 K and up to a maximum pressure of 20 MPa. Fig. 1
illustrates the recorded data as coordinates in a p, T-diagram
together with the saturation curve calculated using the GERG-
2008 EoS [26]. The p, T-range where the GERG-2008 EoS can be
applied and the region of technical relevance are also shown
in Fig. 1. To ensure an error-free operation of the device, ni-
trogen as a reference fluid was investigated at selected con-
ditions over the entire operational range [50]. During a
measurement, the pressure was reduced in intervals of 1 MPa
starting from 20 MPa down to 1 MPa for each isotherm. Thirty
repeated measurements were processed into one definite data
point. The closing procedure of each isothermal measuring
series was the determination of the true sinker mass under
vacuum conditions mS0 to detect any misalignment from the
magnetic suspension couplings.
Uncertainty analysis
The evaluation of uncertainties (all given as expanded un-
certainties, i.e., k ¼ 2) followed the same procedure as in
previous studies [42]. The uncertainty in temperature
amounted to less than 4 mK, the uncertainty in pressure
depended on the range covered by the individual transducer.
Eq. (2) gives the uncertainty relation for the high-pressure
(3e20 MPa) transducer, and Eq. (3) for the low-pressure
transducer (0e3 MPa), respectively. The uncertainty in pres-
sure in both mixtures remained below 0.005 MPa.
UðpÞ = MPa ¼ 75$106 $p = MPa þ 3:5$103 (2)
UðpÞ = MPa ¼ 60$106 $p = MPa þ 1:7$103 (3)
The uncertainty of the density value was calculated by
executing the uncertainty propagation law on Eq. (1) accord-
ing to the procedures given in the Guide to the Expression of

Table 4 e Contributions to the expanded overall uncertainty in density UT (rexp) (k ¼ 2) for the two studied natural gas
mixtures in the temperature range from 260 to 350 K.
Source of uncertainty Units Contribution (k ¼ 2) Estimation in density (k ¼ 2)
kg$m3 %
11 M (BAM-G 420) natural gas
BAM no.: C49358-090825
Temperature T K 0.004 <0.008 <0.005
Pressure p MPa 0.005 (0.002e0.069) (0.015e0.191)
Density r kg$m3 (0.023e0.049) (0.023e0.049) (0.022e0.378)
Composition xi mol$mol1 <0.0004 <0.023 (0.005e0.029)
Overall uncertainty UT (rexp) (0.024e0.081) (0.031e0.507)
H2-enriched natural gas
BAM no.: 8099-160905
Temperature T K 0.004 <0.007 <0.004
Pressure p MPa 0.005 (0.002e0.049) (0.006e0.189)
Density r kg$m3 (0.024e0.049) (0.024e0.049) (0.022e0.372)
Composition xi mol$mol1 <0.0004 <0.007 <0.003
Overall uncertainty UT (rexp) (0.024e0.071) (0.029e0.502)
21990 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8

Table 5 e Experimental (p, r, T) measurements for the 11 M (BAM-G 420) natural gas, relative and absolute expanded
uncertainty in density (k ¼ 2) U (rexp), and relative deviations from the GERG-2008 and AGA8-DC92 EoS; where T is the
temperature (ITS-90), p the pressure, rexp the experimental density, and rGERG and rAGA the densities calculated from the
GERG-2008 and the AGA8-DC92 EoS.
T/Ka p/MPaa rexp/kg$m3 U (rexp)/kg$m3 (k ¼ 2) 102 U (rexp)/rexp 102 (rexperGERG)/rGERG 102 (rexperAGA)/rAGA
259.989 17.977 225.658 0.049 0.022 0.065 0.127
259.992 17.023 216.671 0.048 0.022 0.074 0.119
259.994 16.014 206.202 0.047 0.023 0.084 0.109
259.992 15.025 194.876 0.045 0.023 0.092 0.100
259.994 14.012 182.092 0.044 0.024 0.095 0.093
259.995 13.021 168.450 0.042 0.025 0.090 0.093
259.994 12.009 153.491 0.041 0.027 0.075 0.099
259.994 11.014 138.040 0.039 0.028 0.047 0.098
259.996 10.013 122.146 0.037 0.030 0.025 0.099
259.990 9.008 106.300 0.035 0.033 0.007 0.096
259.993 7.988 90.764 0.034 0.037 0.010 e
259.991 7.000 76.503 0.032 0.042 0.020 0.077
259.992 5.999 62.980 0.030 0.048 0.019 0.075
259.992 4.998 50.417 0.029 0.057 0.015 0.072
259.993 4.002 38.839 0.028 0.071 0.008 0.068
259.994 2.997 28.021 0.026 0.094 0.005 0.059
259.992 1.992 17.972 0.025 0.140 0.001 0.048
259.991 0.998 8.701 0.024 0.278 0.013 0.041
274.977 19.622 213.256 0.048 0.022 0.057 0.117
274.977 19.015 208.165 0.047 0.023 0.061 0.114
274.975 17.999 199.103 0.046 0.023 0.067 0.108
274.976 17.010 189.597 0.045 0.024 0.071 0.102
274.976 16.002 179.179 0.044 0.024 0.072 0.099
274.973 14.994 168.038 0.042 0.025 0.071 0.099
274.971 14.005 156.441 0.041 0.026 0.065 0.100
274.972 13.007 144.175 0.040 0.027 0.056 0.102
274.972 12.009 131.463 0.038 0.029 0.044 0.100
274.971 11.007 118.489 0.037 0.031 0.027 0.092
274.972 10.004 105.462 0.035 0.033 0.018 0.090
274.971 9.001 92.621 0.034 0.036 0.010 0.087
274.971 8.000 80.158 0.032 0.040 0.005 0.085
274.971 7.001 68.196 0.031 0.045 0.001 0.079
274.972 5.998 56.751 0.030 0.052 0.003 0.075
274.972 4.997 45.910 0.028 0.062 0.002 0.070
274.972 4.001 35.703 0.027 0.076 0.001 0.064
274.971 3.000 26.006 0.026 0.101 0.003 0.054
274.968 2.000 16.860 0.025 0.149 <0.001 0.038
274.969 0.998 8.183 0.024 0.295 0.001 0.021
299.934 16.303 150.195 0.040 0.027 0.027 0.072
299.934 16.008 147.459 0.040 0.027 0.026 0.072
299.933 15.003 137.934 0.039 0.028 0.022 0.073
299.932 13.999 128.156 0.038 0.030 0.017 0.073
299.933 13.000 118.212 0.037 0.031 0.012 0.072
299.934 12.000 108.142 0.036 0.033 0.008 0.072
299.932 11.001 98.044 0.034 0.035 0.001 0.065
299.931 10.001 87.967 0.033 0.038 0.002 0.066
299.930 9.001 78.002 0.032 0.041 0.003 0.065
299.931 7.999 68.196 0.031 0.045 0.004 0.064
299.930 6.998 58.628 0.030 0.051 0.004 0.060
299.926 5.996 49.322 0.029 0.058 0.004 0.056
299.927 4.998 40.335 0.028 0.069 0.004 0.049
299.926 3.998 31.645 0.027 0.085 0.001 0.044
299.928 2.997 23.266 0.026 0.111 0.003 0.033
299.926 1.997 15.204 0.025 0.164 0.002 0.025
299.923 0.997 7.446 0.024 0.323 0.005 0.021
324.932 18.569 146.171 0.040 0.027 0.047 0.099
324.934 18.000 141.876 0.039 0.028 0.046 0.100
324.936 16.998 134.140 0.038 0.029 0.044 0.102
324.936 15.997 126.218 0.038 0.030 0.042 0.104
324.936 14.997 118.150 0.037 0.031 0.038 0.103
324.937 13.998 109.947 0.036 0.033 0.036 0.102
324.937 12.999 101.654 0.035 0.034 0.034 0.101
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8 21991

Table 5 e (continued )
T/Ka p/MPaa rexp/kg$m3 U (rexp)/kg$m3 (k ¼ 2) 102 U (rexp)/rexp 102 (rexperGERG)/rGERG 102 (rexperAGA)/rAGA
324.937 11.998 93.287 0.034 0.036 0.031 0.099
324.936 10.997 84.908 0.033 0.039 0.024 0.090
324.936 10.001 76.578 0.032 0.042 0.027 0.091
324.935 8.998 68.248 0.031 0.045 0.025 0.087
324.933 7.997 60.023 0.030 0.050 0.023 0.082
324.933 6.997 51.928 0.029 0.056 0.021 0.075
324.934 5.997 43.970 0.028 0.064 0.022 0.070
324.934 4.997 36.174 0.027 0.076 0.023 0.065
324.934 3.997 28.555 0.026 0.093 0.022 0.057
324.934 2.997 21.122 0.026 0.121 0.023 0.051
324.930 1.990 13.831 0.025 0.179 0.023 0.044
324.929 0.998 6.837 0.024 0.351 0.032 0.044
349.919 19.701 136.343 0.039 0.028 0.031 0.095
349.918 18.999 131.749 0.038 0.029 0.030 0.096
349.918 17.997 125.065 0.037 0.030 0.029 0.098
349.919 16.995 118.246 0.037 0.031 0.027 0.098
349.919 15.997 111.336 0.036 0.032 0.026 0.098
349.921 14.996 104.305 0.035 0.034 0.024 0.096
349.920 13.996 97.202 0.034 0.035 0.023 0.094
349.919 12.995 90.022 0.033 0.037 0.022 0.092
349.919 11.997 82.823 0.033 0.039 0.021 0.089
349.918 10.999 75.612 0.032 0.042 0.018 0.082
349.919 9.997 68.368 0.031 0.045 0.019 0.079
349.919 8.996 61.155 0.030 0.049 0.020 0.076
349.919 7.996 53.989 0.029 0.054 0.021 0.072
349.918 6.997 46.892 0.029 0.061 0.019 0.065
349.918 5.996 39.864 0.028 0.070 0.021 0.061
349.918 4.997 32.935 0.027 0.082 0.022 0.057
349.918 3.996 26.095 0.026 0.100 0.025 0.055
349.918 2.998 19.395 0.025 0.131 0.026 0.050
349.917 1.997 12.791 0.025 0.193 0.020 0.037
349.917 0.997 6.323 0.024 0.378 0.031 0.042
a
Expanded uncertainties (k ¼ 2) in temperature and pressure are U(T) ¼ 0.004 K and U(p) ¼ 0.005 MPa, respectively.

Uncertainty in Measurement (GUM) [51]. From Eq. (1), the true
sinker mass mS0, the apparent sinker mass in the medium mSf,
and the sinker volume VS (T, p) contribute to the uncertainty.
Additionally, the uncertainty of the apparent sinker mass mSf
includes the entry from calibration, resolution, repeatability,
and balance drift. Since the sinker volume VS is affected by
temperature and pressure, Eq. (4) describes the effect of
thermal expansion and pressure distortion as a function of
density.
UðrÞ = kg$ m3 ¼ 1:1$104 $r=kg$m3 þ 2:3$102
The overall expanded uncertainty in density UT (r) includes
the uncertainties of density, temperature, pressure, and ulti-
mately the composition of the gas mixture, see Eq. (5).
calculated from GERG-2008 EoSe via REFPROP [37]. The indi-
vidual uncertainty contributions are summarized in Table 4.
Results and discussion
The experimental volumetric (p, r, T) data measured for the
two natural gas mixtures during this work are given in Table 5
for the 11 M (BAM-G 420) mixture and Table 6 for the H2-
enriched mixture together with the corresponding expanded
(4) uncertainty in density from Eq. (4) and expressed both in
density units (kg m3, i.e., absolute value) and as percentage of
the measured density (i.e., relative value). The experimental
data were compared to the corresponding density data
calculated from the GERG-2008 and AGA8-DC92 EoS. Two

2 !2 !2 !2 30:5
    X  vr 
vr vr
UT ðrÞ ¼ 2$4uðrÞ þ 2
$uðpÞ þ $uðTÞ þ $uðxi Þ 5 (5)
vp T;x vT p;x i
vxi T;p;xj sxi

In Eq. (5), apart from temperature T and pressure p, xi is the columns in the Tables 5 and 6 represent the relative deviations
amount-of-substance (mole) fraction of each individual between experimental and calculated data and the corre-
mixture component. Partial derivatives of Eq. (5) may be sponding data are plotted in Figs. 2e5.
21992 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8

Table 6 e Experimental (p, r, T) measurements for the H2-enriched natural gas mixture, relative and absolute expanded
uncertainty in density (k ¼ 2) U (rexp), and relative deviations from the GERG-2008 and AGA8-DC92 EoS; where T is the
temperature (ITS-90), p the pressure, rexp the experimental density, and rGERG and rAGA the densities calculated from the
GERG-2008 and the AGA8-DC92 EoS.
T/Ka p/MPaa rexp/kg m3 U (rexp)/kg$m3 (k ¼ 2) 102 U (rexp)/rexp 102 (rexperGERG)/rGERG 102 (rexperAGA)/rAGA
260.041 19.843 224.730 0.049 0.022 0.254 0.193
260.038 19.011 217.428 0.048 0.022 0.268 0.178
260.040 18.010 208.053 0.047 0.023 0.280 0.156
260.041 17.008 197.993 0.046 0.023 0.288 0.133
260.040 16.010 187.277 0.045 0.024 0.291 0.114
260.040 15.012 175.878 0.043 0.025 0.285 0.097
260.040 14.011 163.801 0.042 0.026 0.269 0.083
260.039 13.012 151.191 0.040 0.027 0.245 0.071
260.038 12.011 138.130 0.039 0.028 0.214 0.063
260.039 11.009 124.803 0.037 0.030 0.171 0.051
260.041 10.008 111.413 0.036 0.032 0.135 0.050
260.040 9.005 98.132 0.034 0.035 0.104 e
260.041 8.010 85.247 0.033 0.039 0.075 0.059
260.040 7.007 72.690 0.031 0.043 0.050 0.060
260.040 6.001 60.600 0.030 0.050 0.035 0.064
260.038 4.999 49.124 0.029 0.059 0.025 0.064
260.039 4.001 38.246 0.028 0.072 0.018 0.060
260.040 2.993 27.840 0.026 0.095 <0.001 0.039
260.041 1.998 18.098 0.025 0.140 0.013 0.045
260.041 0.999 8.815 0.024 0.275 0.001 0.019
275.020 19.234 196.179 0.046 0.023 0.225 0.126
275.019 18.008 185.120 0.044 0.024 0.228 0.111
275.018 17.013 175.624 0.043 0.025 0.226 0.101
275.018 16.006 165.559 0.042 0.025 0.220 0.092
275.016 15.004 155.107 0.041 0.026 0.209 0.084
275.017 14.008 144.361 0.040 0.027 0.193 0.076
275.016 13.007 133.256 0.038 0.029 0.174 0.070
275.015 12.006 121.941 0.037 0.030 0.155 0.067
275.015 11.004 110.531 0.036 0.032 0.128 0.060
275.013 10.006 99.141 0.035 0.035 0.111 0.063
275.014 9.003 87.821 0.033 0.038 0.092 0.065
275.013 8.002 76.705 0.032 0.042 0.076 0.069
275.013 7.001 65.854 0.031 0.047 0.060 0.068
275.013 6.000 55.328 0.030 0.053 0.049 0.067
275.013 5.003 45.205 0.028 0.063 0.038 0.062
275.012 4.006 35.445 0.027 0.077 0.027 0.054
275.011 3.001 26.000 0.026 0.101 0.012 0.037
275.011 2.000 16.957 0.025 0.148 0.003 0.025
275.011 0.999 8.296 0.024 0.291 0.009 0.005
299.961 19.851 171.675 0.043 0.025 0.162 0.089
299.961 19.009 165.029 0.042 0.025 0.161 0.086
299.960 18.003 156.843 0.041 0.026 0.157 0.084
299.960 17.002 148.434 0.040 0.027 0.152 0.081
299.960 16.003 139.800 0.039 0.028 0.145 0.078
299.960 15.002 130.945 0.038 0.029 0.137 0.075
299.960 14.002 121.924 0.037 0.030 0.127 0.072
299.960 13.003 112.776 0.036 0.032 0.116 0.069
299.958 12.003 103.536 0.035 0.034 0.108 0.069
299.958 11.000 94.230 0.034 0.036 0.091 0.061
299.958 10.002 84.966 0.033 0.039 0.082 0.062
299.958 9.001 75.735 0.032 0.042 0.071 0.059
299.957 8.000 66.591 0.031 0.046 0.061 0.057
299.956 7.002 57.603 0.030 0.052 0.049 0.052
299.956 5.999 48.731 0.029 0.059 0.042 0.049
299.957 4.999 40.073 0.028 0.069 0.031 0.042
299.959 3.998 31.613 0.027 0.085 0.021 0.033
299.958 2.991 23.309 0.026 0.111 0.008 0.020
299.957 1.999 15.356 0.025 0.162 0.004 0.016
299.958 0.999 7.557 0.024 0.318 0.001 0.007
324.963 19.928 150.681 0.040 0.027 0.107 0.064
324.962 19.001 144.143 0.040 0.027 0.105 0.065
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8 21993

Table 6 e (continued )
T/Ka p/MPaa rexp/kg m3 U (rexp)/kg$m3 (k ¼ 2) 102 U (rexp)/rexp 102 (rexperGERG)/rGERG 102 (rexperAGA)/rAGA
324.963 17.999 136.914 0.039 0.028 0.102 0.064
324.963 17.000 129.541 0.038 0.029 0.098 0.063
324.964 16.000 122.027 0.037 0.030 0.094 0.061
324.963 14.998 114.372 0.036 0.032 0.088 0.059
324.963 14.001 106.647 0.035 0.033 0.082 0.056
324.963 13.002 98.826 0.034 0.035 0.076 0.054
324.962 12.001 90.934 0.034 0.037 0.069 0.051
324.963 11.000 83.015 0.033 0.039 0.057 0.042
324.963 9.999 75.089 0.032 0.042 0.053 0.042
324.962 8.999 67.186 0.031 0.046 0.043 0.036
324.962 7.999 59.321 0.030 0.051 0.037 0.033
324.962 6.999 51.523 0.029 0.056 0.028 0.027
324.962 5.998 43.796 0.028 0.064 0.023 0.025
324.961 4.999 36.180 0.027 0.076 0.018 0.022
324.962 3.998 28.673 0.026 0.092 0.012 0.018
324.961 2.985 21.199 0.026 0.121 0.002 0.009
324.961 1.999 14.050 0.025 0.177 0.001 0.008
324.962 0.998 6.945 0.024 0.345 0.005 0.001
349.946 19.941 134.571 0.039 0.029 0.110 0.091
349.946 19.000 128.613 0.038 0.029 0.108 0.090
349.945 17.997 122.149 0.037 0.030 0.106 0.089
349.946 16.995 115.580 0.036 0.031 0.102 0.087
349.944 15.998 108.944 0.036 0.033 0.099 0.085
349.946 14.996 102.190 0.035 0.034 0.094 0.081
349.946 13.998 95.389 0.034 0.036 0.089 0.077
349.945 12.998 88.509 0.033 0.038 0.085 0.074
349.944 11.991 81.546 0.032 0.040 0.073 0.063
349.945 10.996 74.624 0.032 0.042 0.064 0.056
349.945 9.997 67.658 0.031 0.046 0.062 0.055
349.946 8.998 60.691 0.030 0.050 0.057 0.052
349.947 7.998 53.731 0.029 0.055 0.052 0.049
349.947 6.997 46.792 0.029 0.061 0.045 0.044
349.947 5.998 39.899 0.028 0.070 0.041 0.042
349.946 4.998 33.064 0.027 0.082 0.037 0.040
349.946 3.998 26.284 0.026 0.100 0.032 0.037
349.946 2.985 19.492 0.025 0.130 0.026 0.031
349.946 1.999 12.964 0.025 0.190 0.025 0.031
349.946 0.999 6.431 0.024 0.372 0.007 0.012
a
Expanded uncertainties (k ¼ 2) in temperature and pressure are U(T) ¼ 0.004 K and U(p) ¼ 0.005 MPa, respectively.

Fig. 2 e Relative deviations in density of experimental (p, r,
T) data of the 11 M natural gas-like mixture rexp from
density values calculated from the GERG-2008 EoS rGERG
versus pressure p: ,, T ¼ 260 K; ⋄, T ¼ 275 K; △, T ¼ 300 K;
✕, T ¼ 325 K; B, T ¼ 350 K. Error bars on the 260-K
isotherm indicate the expanded uncertainty (k ¼ 2) of the
experimental density data calculated with Eq. (4).
Fig. 3 e Relative deviations in density of experimental (p, r,
T) data of the 11 M natural gas-like mixture rexp from
density values calculated from the AGA8-DC92 EoS rAGA
versus pressure p: ,, T ¼ 260 K; ⋄, T ¼ 275 K; △, T ¼ 300 K;
✕, T ¼ 325 K; B, T ¼ 350 K. Error bars on the 260-K
isotherm indicate the expanded uncertainty (k ¼ 2) of the
experimental density data calculated with Eq. (4).
21994 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8
Fig. 4 e Relative deviations in density of experimental (p, r,
T) data of the H2-enriched natural gas mixture rexp from
density values calculated from the GERG-2008 EoS rGERG
versus pressure p: ,, T ¼ 260 K; ⋄, T ¼ 275 K; △, T ¼ 300 K;
✕, T ¼ 325 K; B, T ¼ 350 K. Error bars on the 260-K
isotherm indicate the expanded uncertainty (k ¼ 2) of the
experimental density data calculated with Eq. (4).
Figs. 2 and 3 show the relative deviations of experimental
density data (rexp) from density data calculated by the GERG-
2008 (rGERG) and AGA8-DC92 (rAGA) EoS versus pressure p at a
constant temperature T for the 11 M natural gas mixture. Figs.
4 and 5 represent the equivalent for the H2-enriched natural
gas mixture.
The intrinsic uncertainty of the GERG-2008 EoS in the gas-
phase region over the temperature range from 250 to 450 K
and at pressures up to 35 MPa amounts to 0.1% in density [26].
A look at the Figs. 2e5 shows that a negative deviation of
the calculated density compared to the experimental density
dominates for both gas mixtures and both equation-of-state
models. The relative deviations approximate to zero towards
low pressures as it is expected from the ideal gas limit and is a
Fig. 5 e Relative deviations in density of experimental (p, r,
T) data of the H2-enriched natural gas mixture rexp from
density values calculated from the AGA8-DC92 EoS rAGA
versus pressure p: ,, T ¼ 260 K; ⋄, T ¼ 275 K; △, T ¼ 300 K;
✕, T ¼ 325 K; B, T ¼ 350 K. Error bars on the 260-K
isotherm indicate the expanded uncertainty (k ¼ 2) of the
experimental density data calculated with Eq. (4).
prove of the stated composition of the mixtures and of its
stability. For the 11 M mixture processed by the GERG-2008 EoS
shown in Fig. 2, the data follow a relatively flat sinusoid curve
but all data remain within the 0.1-% margin, with the largest
deviation for the lowest temperature T ¼ 260 K. The effect
becomes stronger and more temperature dependent at pres-
sures p > 10 MPa. The results for AGA8-DC92 on the 11 M
mixture in Fig. 3 look different. The deviations are all negative
and the deviation increases slowly but monotonously towards
increasing pressure. A comparison of the performance of both
equation-of-state models shows that GERG-2008 performs
better on 11 M than AGA8-DC92, but the gain diminishes when
going towards lower temperatures. The results for the H2-
enriched gas mixtures are apparently different. The GERG-
2008 EoS shown in Fig. 4 resulted in a quite similar pattern
as for the 11 M mixture, but more strongly pronounced. The
deviation becomes steeper at about 5 MPa and goes through a
maximum of about 0.30% at around 16 MPa. There is a visible
temperature dependence. In contrast, the AGA8-DC92, shown
in Fig. 5, shows a smaller deviation and a less pronounced
temperature dependence until a pressure of 15 MPa where a
steeper deviation begins. As expected, the performance on the
H2-enriched mixture is generally poorer and, surprisingly, the
AGA8-DC92 gives better results.
The findings illustrated in Figs. 2e5 were further evalu-
ated by statistical parameters that were already applied in
previous studies [50] and are given in Table 7. Eq. (6) defines
the average absolute deviation AAD, Eq. (7) the so-called Bias
that quantifies the average deviation of the data set, and Eq.
(8) represents the root mean square RMS. The subscript “EoS”
is replaced by “GERG” or “AGA” to denote the applied
equation-of-state model in the corresponding Tables and
Figures.
n  
1X r
102 $ i;exp
 ri;EoS 

AAD ¼   (6)
n i¼1 ri;EoS
n  
1X ri;exp  ri;EoS
Bias ¼ 102 $ (7)
n i¼1 ri;EoS
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u n  
u1X ri;exp  ri;EoS 2
RMS ¼ t 102 $ (8)
n i¼1 ri;EoS
Additionally, MaxD stands for the maximum relative de-
viation in the considered data set given as absolute value. For
the 11 M natural gas the AAD of 0.027 from the GERG-2008 EoS
was lower than the corresponding value of 0.078 from the
AGA8-DC92 EoS. A similar relation was found for the RMS
values, and e expectedly e GERG-2008 also produced a lower
value for MaxD. However, the result for the H2-enriched nat-
ural gas was the opposite. Here, the application of AGA8-DC92
gave a lower AAD value of 0.062 compared to 0.095 for GERG-
2008. Consequently, AGA8-DC92 produced lower values also
for RMS and MaxD, respectively.
The statistical analysis according to the Eqs. (6)e(8) was
applied on some selected recently published literature data
that had executed both equation-of-state models on their re-
sults for comparison [30e32].
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8 21995

Table 7 e Statistical parameters of the (p, r, T) data set with respect to the GERG-2008 and AGA8-DC92 EoS for the two
studied natural gas mixtures including literature data for comparable mixtures.
Reference Identifier Covered ranges N Experimental vs. GERG-2008 Experimental vs. AGA8-DC92
T/K p/MPa AAD Bias RMS MaxD/% AAD Bias RMS MaxD/%
This work (2018) 11 M natural gas 260e350 1e20 94 0.027 0.025 0.036 0.095 0.078 0.078 0.081 0.127
This work (2018) H2-enriched 260e350 1e20 99 0.095 0.095 0.123 0.291 0.062 0.062 0.070 0.193
natural gas
Richter et al. (2014) [30] NG1 273e293 1.0e8.0 37 0.021 0.018 0.023 0.0375 0.010 0.001 0.013 0.0338
Richter et al. (2014) [30] NG2 273e293 1.0e8.0 36 0.032 0.032 0.036 0.0664 0.013 0.004 0.018 0.0509
Richter et al. (2014) [30] NG3 283 1.0e8.0 13 0.014 0.009 0.015 0.0256 0.027 0.026 0.031 0.0447
Atilhan et al. (2015) [31] M88C1 270e340 3.5e34.5 32 0.221 0.045 0.261 0.639 0.365 0.299 0.384 0.610
Atilhan et al. (2015) [31] M94C1 270e340 3.5e34.5 61 0.186 0.144 0.215 0.516 0.094 0.039 0.112 0.361
Ahmadi et al. (2017) [32] natural gas 323e415 1.3e58.4 110 0.135 0.002 0.303 2.18 0.1a
a
Reference 32 does not provide tabled individual data of the deviation resulting from the AGA8-DC92 EoS.

The mixtures studied by Richter and co-workers were
H2-enriched, 21-component high-calorific gases with 5
(NG1), 10 (NG2), and 30 mol-% hydrogen (NG3) originating
from a pipeline. The characteristic deviation pattern found
in their study looks different to our results. Data processing
with GERG-2008 resulted mostly in a positive deviation that
goes through a flat maximum between 4 and 6 MPa for all
three mixtures. The AGA8-DC92 EoS produced more nega-
tive deviations for the respective coordinates, but the values
for AAD, Bias, and RMS turned out to be smaller than the
corresponding values for the GERG-2008 EoS except for NG3,
i.e., the mixture with the highest hydrogen content of
30 mol-%.
The mixtures studied by Atilhan et al. contained heavier
alkanes but no hydrogen [31]. The methane content of M88C1
(88 mol-%) was smaller than for M94C1 (94 mol-%) with that
difference being merely compensated by higher amounts of
ethane (5.8 mol-% for M88C1 and 1.9 mol-% for M94C1) and
propane (3.3 mol-% for M88C1 and 1.8 mol-% for M94C1),
respectively. In both mixtures, carbon dioxide (1.5 mol-%) and
nitrogen (2.5 mol-%) were at approximately similar concen-
trations. Data analysis with the GERG-2008 EoS resulted in a
sinusoid function of both positive (at p > 25 MPa) and negative
(at p < 25 MPa) deviations, with no significant temperature
dependence for M88C1. The corresponding diagram of M94C1,
however, shows mostly negative deviations that turn into
positive values not before p > 30 MPa. A temperature depen-
dence clearly manifested in the largest deviation at low tem-
peratures. Data processing of M88C1 with the AGA8-DC92 EoS
gave a similar pattern but with a stronger temperature
dependence than for GERG-2008; the results for M94C1 display
a good coincidence with minor deviations only. An assess-
ment of the statistical parameters given in Table 7 states a
better performance of GERG-2008 on M88C1 and of AGA8-
DC92 on M94C1. However, the difference between the two
models remains rather small.
Ultimately, Ahmadi et al. investigated a rather “ordinary”
natural gas mixture of 7 compounds (methane: 88 mol-%,
ethane: 6.0 mol-%, propane: 2.0 mol-%, n-butane: 0.3 mol-%,
isobutane: 0.2 mol-%, carbon dioxide: 2.0 mol-%, nitrogen:
1.5 mol-%) [32]. A characteristic feature of their results (the
detailed plot was provided for the GERG-2008 EoS only) is that
scattering and (significant) deviation starts at about pressures
of 15 MPa and further increases towards lower pressure.
When this area is not considered in the analysis, experimental
and calculated data show good coincidence even at the
highest pressures investigated.
Conclusions
In this work, new high-precision experimental (p, r, T) data
for two multicomponent natural gas mixtures were recor-
ded. The gas mixtures for this study mimic real natural gas
mixtures, one being a conventional 11-component gas, the
other a 13-component H2-enriched mixture proposed by
CCQM. They were prepared by gravimetry to create a met-
rologically traceable mixture which qualifies as reference
material. The experiments were done using a single-sinker
densimeter with magnetic suspension coupling in a tem-
perature range between 260 and 350 K and up to pressures
of 20 MPa. Subsequently, the new data were compared with
the corresponding densities calculated with the two estab-
lished reference EoS for natural gases, namely GERG-2008
and AGA8-DC92. The actual performance of the equation-
of-state model can be tested using consolidated data of
real mixtures. The application of the two models on the
density data from the present study resulted in an average
absolute deviation that remained below 0.1% for both nat-
ural gas mixtures and both models, respectively, for the p,
T-range investigated.
For the 11 M natural gas mixture processed by the GERG-2008
EoS all data remain within the 0.1-% margin, with the largest
deviation for the lowest temperature T ¼ 260 K. The results for
AGA8-DC92 on the 11 M mixture remain also within the 0.1-%
margin, except at higher pressures, where the differences with
the EoS are slightly bigger than 0.1%. A comparison of the per-
formance of both equation-of-state models shows that GERG-
2008 performs better on 11 M than AGA8-DC92.
The results for the H2-enriched gas mixtures are different.
The deviations of experimental data from the GERG-2008 EoS
are not within the 0.1-% margin at high pressures, this
pressure limit being lower for lower temperatures (17 MPa for
350 K, and only 9 MPa for 260 K). The deviation may be as high
as 0.30% at 260 K and 16 MPa. In contrast, the deviations from
the AGA8-DC92 EoS are smaller and are found outside the 0.1
21996 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8
%-margin only for the two lower temperatures measured
(i.e., 260 K and 275 K) and pressures above 16 MPa. The
maximum deviation is near 0.2% at 260 K and 20 MPa. In this
case, the performance of AGA8-DC92 is better than that of
GERG 2008.
Generally speaking, the performance of the AGA8-DC92
EoS is satisfactory when dealing both with natural gas mix-
tures and with hydrogen-enriched natural gas mixtures, giv-
ing in both cases similar results. Otherwise, the GERG-2008
EoS displays a better performance than AGA8-DC92 when
applied on natural gas-type mixtures, but presents higher
deviations for hydrogen-enriched natural gas mixtures at low
temperatures and high pressures.
Nevertheless, our results only cover a limited p, T-sector
compared to the approved overall operational range of both
equation-of-state models. As the literature comparison
shows, a different behavior can occur at other temperatures
and pressures, which neither can directly be drawn from an
extrapolation nor is easily predictable. Improvement of the
theoretical models designed for natural gas, especially for
their use with hydrogen-enriched natural gas mixtures, will
need e consolidated e experimental thermodynamic proper-
ties of binary mixtures of hydrogen with the main compo-
nents of natural gas [50, 52]. The user community needs to
know about the generally accepted limits of the models and
their performance in the vicinity of these limits for different
mixtures. Our study was done at conditions that are often
encountered during applications, and the density data ob-
tained from samples of metrological traceability were pro-
cessed with both reference EoS, which produced results of
good quality.
Acknowledgments
Support for this work came from the projects ‘Caracter-
izacion de gases energeticos sostenibles (biogas e hidrogeno),
producidos con recursos renovables bioma
sicos y eo

licos,
para su incorporacio
n a la red de gas natural’, ENE2017-
88474-R, funded by the Spanish Government and
‘Revalorizacio
n de recursos renovables regionales bio-

sicos y eo

licos para la produccio

n de gases energe
ticos
ma




sostenibles (biogas e hidrogeno) y su incorporacion en la red
de gas natural’, VA035U16 and ‘UIC-114’ from the Junta de

n.
Castilla y Leo
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ijhydene.2018.10.027.
references
[1] De Falco M, Basile A, editors. Enriched methane. The first
step towards the hydrogen economy. Springer International
Publishing Switzerland; 2016. https://doi.org/10.1007/978-3-
319-22192-2.
[2] Ball M, Wietschel M, editors. The hydrogen economy.
Opportunities and Challenges. Cambridge, UK: Cambridge
University Press; 2009.
[3] Bica

kova

O, Straka P. Production of hydrogen from
renewable resources and its effectiveness. Int J Hydrogen
Energy 2012;37:11563e78. https://doi.org/10.1016/j.ijhydene.
2012.05.047.
[4] Granovskii M, Dincer I, Rosen MA. Greenhouse gas emissions
reduction by use of wind and solar energies for hydrogen and
electricity production: economic factors. Int J Hydrogen
Energy 2007;32:927e31. https://doi.org/10.1016/j.ijhydene.
2006.09.029.
[5] Balat H, Kırtay E. Hydrogen from biomass e present scenario
and future prospects. Int J Hydrogen Energy 2010;35:7416e26.
https://doi.org/10.1016/j.ijhydene.2010.04.137.
[6] Acar C, Dincer I. Comparative assessment of hydrogen
production methods from renewable and non-renewable
sources. Int J Hydrogen Energy 2014;39:1e12. https://doi.org/
10.1016/j.ijhydene.2013.10.060.
[7] Dincer I, Acar C. Innovation in hydrogen production. Int J
Hydrogen Energy 2017;42:14843e64. https://doi.org/10.1016/
j.ijhydene.2017.04.107.
[8] Schiebahn S, Grube T, Robinius M, Tietze V, Kumar B,
Stolten D. Power to gas: technological overview, systems
analysis and economic assessment for a case study in
Germany. Int J Hydrogen Energy 2015;40:4285e94. https://
doi.org/10.1016/j.ijhydene.2015.01.123.
[9] AlRafea K, Fowler M, Elkamel A, Hajimiragha A. Integration
of renewable energy sources into combined cycle power
plants through electrolysis generated hydrogen in a new
designed energy hub. Int J Hydrogen Energy 2016;41:
16718e28. https://doi.org/10.1016/j.ijhydene.2016.06.256.
[10] Gahleitner G. Hydrogen from renewable electricity: an
international review of power-to-gas pilot plants for
stationary applications. Int J Hydrogen Energy 2013;38:
2039e61. https://doi.org/10.1016/j.ijhydene.2012.12.010.
[11] Polman E, de Laat H, Stappenbeit J, Peereboom P,
Bouwman W, de Bruin B, et al. Reduction of CO2 emissions by
adding hydrogen to natural gas. Cheltenham, UK: IEA
Greenhouse Gas R&D Programme; 2003. Report No. PH 4/24.
[12] Tabak J. Natural gas and hydrogen. New York NY: Facts On
File; 2009.
[13] Melaina MW, Antonia O, Penev M. Blending hydrogen into
natural gas pipeline networks: a review of key issues. Golden
CO: National Renewable Energy Laboratory (NREL); 2013.
Technical Report NREL/TP-5600-51995, http://www.nrel.gov/
docs/fy13osti/51995.pdf. [Accessed 7 December 2017].
[14] Messaoudani ZL, Rigas F, Binti Hamid MD, Che Hassan CR.
Hazards, safety and knowledge gaps on hydrogen
transmission via natural gas grid: a critical review. Int J
Hydrogen Energy 2016;41:17511e25. https://doi.org/10.1016/
j.ijhydene.2016.07.171.
[15] Schouten JA, Michels JPJ, Janssen-van Rosmalen R. Effect of
H2-injection on the thermodynamic and transportation
properties of natural gas. Int J Hydrogen Energy 2004;29:
1173e80. https://doi.org/10.1016/j.ijhydene.2003.11.003.
[16] Capocelli M, De Falco M. Enriched methane: a ready solution
for the transition towards the hydrogen economy. In: De
Falco M, Basile A, editors. Enriched methane. The first step
towards the hydrogen economy. Springer International
Publishing Switzerland; 2016. p. 1e21. https://doi.org/
10.1007/978-3-319-22192-2_1.
[17] Li T, Hampp F, Lindstedt RP. Experimental study of turbulent
explosions in hydrogen enriched syngas related fuels.
Process Saf Environ Protect 2018;116:663e76. https://doi.org/
10.1016/j.psep.2018.03.032.
[18] Lowesmith BJ, Mumby C, Hankinson G, Puttock JS. Vented
confined explosions involving methane/hydrogen mixtures.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8
Int J Hydrogen Energy 2011;36:2337e43. https://doi.org/
10.1016/j.ijhydene.2012.02.084.
[19] Maroufmashat A, Fowler M. Transition of future energy
system infrastructure; through power-to-gas pathways.
Energies 2017;10:1089. https://doi.org/10.3390/en10081089.
[20] van der Veen AMH, Ziel PR, Li J. Validation of ISO 6974 for the
measurement of the composition of hydrogen-enriched
natural gas. Int J Hydrogen Energy 2015;40:15877e84. https://
doi.org/10.1016/j.ijhydene.2015.10.005.
[21] Mehra RK, Duan H, Juknelevic ius R, Ma F, Li J. Progress in
hydrogen enriched compressed natural gas (HCNG) internal
combustion engines e a comprehensive review. Renew
Sustain Energy Rev 2017;80:1458e98. https://doi.org/10.1016/
j.rser.2017.05.061.
[22] _
Ilbas‚ M, Karyeyen S. A numerical study on combustion
behaviours of hydrogen-enriched low calorific value coal
gases. Int J Hydrogen Energy 2015;40:15218e26. https://
doi.org/10.1016/j.ijhydene.2015.04.141.
[23] _
Karyeyen S, Ilbas ‚ M. Experimental and numerical analysis of
turbulent premixed combustion of low calorific value coal
gases in a generated premixed burner. Fuel 2018;220:586e98.
https://doi.org/10.1016/j.fuel.2018.02.052.
[24] Hora TS, Agarwal AK. Experimental study of the composition
of hydrogen enriched compressed natural gas on engine
performance, combustion and emission characteristics. Fuel
2015;160:470e8. https://doi.org/10.1016/j.fuel.2015.07.078.
[25] Cellek MS, Pınarbas‚ı A. Investigations on performance and
emission characteristics of an industrial low swirl burner
while burning natural gas, methane, hydrogen-enriched
natural gas and hydrogen as fuels. Int J Hydrogen Energy
2018;43:1194e207. https://doi.org/10.1016/j.ijhydene.2017.
05.107.
[26] Kunz O, Wagner W. The GERG-2008 wide-range equation of
state for natural gases and other mixtures: an expansion of
GERG-2004. J Chem Eng Data 2012;57:3032e91. https://
doi.org/10.1021/je300655b.
[27] ISO 20765-2. Natural gas e calculation of thermodynamic
properties e Part 2: single-phase properties (gas, liquid, and
dense fluid) for extended ranges of application. Geneva:
International Organization for Standardization; 2015.
[28] Starling KE, Savidge JL. Compressibility factors of natural gas
and other related hydrocarbon gases e AGA Transmission
Measurement Committee Report No. 8. 2nd ed. Washington
DC: American Gas Association; 1992.
[29] ISO 6142-1. Gas Analysis e preparation of calibration gas
mixtures e Part 1: gravimetric method for Class I mixtures.
Geneva: International Organization for Standardization;
2015.
[30] Richter M, Ben Souissi MA, Span R, Schley P. Accurate (p, r, T,
x) measurements of hydrogen-enriched natural-gas
mixtures at T ¼ (273.15, 283.15, and 293.15) K with pressures
up to 8 MPa. J Chem Eng Data 2014;59:2021e9. https://doi.org/
10.1021/je500181v.
[31] Atilhan M, Aparicio S, Ejaz S, Zhou J, Al-Marri M, Holste JJ,
et al. Thermodynamic characterization of deepwater natural
gas mixtures with heavy hydrocarbon content at high
pressures. J Chem Thermodyn 2015;82:134e42. https://
doi.org/10.1016/j.jct.2012.04.008.
[32] Ahmadi P, Chapoy A, Tohidi B. Density, speed of sound and
derived thermodynamic properties of a synthetic natural
gas. J Nat Gas Sci Eng 2017;40:249e66. https://doi.org/10.1016/
j.jngse.2017.02.009.
[33] Peng D-Y, Robinson DB. A new two-constant equation of
state. Ind Eng Chem Fund 1976;15:59e64. https://doi.org/
10.1021/i160057a011.
[34] Soave GS. An effective modification of the Benedict-Webb-
Rubin equation of state. Fluid Phase Equil 1999;164:157e72.
https://doi.org/10.1016/S0378-3812(99)00252-6.
21997
[35] PTB Specification PTB-A 7.63. Specification of calibration
gases for analyzers of calorific values and composition of gas
mixtures. Braunschweig, Germany: Physikalisch-Technische
Bundesanstalt; 2011.
[36] CCQM-K 118 Natural Gas; technical protocol V4.0d, CCQM
working group on gas analysis. Se vres, France: Bureau
International des Poids et Mesures BIPM; 2017.
[37] Lemmon EW, Huber ML, McLinden MO. NIST standard
reference database 23: reference fluid thermodynamic and
transport properties-REFPROP, version 9.1. Gaithersburg MD:
National Institute of Standards and Technology, Standard
Reference Data Program; 2013.
[38] ISO 12963. Gas analysis e Comparison methods for the
determination of the composition of gas mixtures based on
one- and two-point calibration. Geneva: International
Organization for Standardization; 2017.
[39] Wagner W, Kleinrahm R. Densimeters for very accurate
density measurements of fluids over large ranges of
temperature, pressure, and density. Metrologia
2004;41:S24e39. https://doi.org/10.1088/0026-1394/41/2/S03.
[40] Brachtha € user K, Kleinrahm R, Lo
€ sch HW, Wagner W.
Entwicklung eines neuen Dichtemeßverfahrens und Aufbau
einer Hochtemperatur-Hochdruck-Dichtemeßanlage,
Fortschr. Ber. VDI, Series 8, no. 173. Germany: VDI-Verlag.
Düsseldorf; 1993.
[41] Klimeck J, Kleinrahm R, Wagner W. An accurate single-
sinker densimeter and measurements of the (p, r, T) relation
of argon and nitrogen in the temperature range from (235 to
520) K at pressures up to 30 MPa. J Chem Thermodyn
1998;30:1571e88. https://doi.org/10.1016/j.jcht.1998.0421.
[42] Monde
jar ME, Segovia JJ, Chamorro CR. Improvement of the
measurement uncertainty of a high accuracy single sinker
densimeter via setup modifications based on a state point
uncertainty analysis. Measurement 2011;44:1768e80. https://
doi.org/10.1016/j.measurement.2011.07.012.
[43] Herna
ndez-Go
mez R, Ferna

ndez-Vicente TE, Martı́n
Gonza
lez MC, Monde
jar ME, Chamorro CR. Integration of
biogas in the natural gas grid: thermodynamic
characterization of a biogas-like mixture. J Chem Thermodyn
2015;84:60e6. https://doi.org/10.1016/j.jct.2014.12.028.
[44] Richter M, Kleinrahm R, Lentner R, Span R. Development of a
special single-sinker densimeter for cryogenic liquid
mixtures and first results for a liquefied natural gas (LNG). J
Chem Thermodyn 2016;93:205e21. https://doi.org/10.1016/
j.jct.2015.09.034.
[45] McLinden MO. Experimental techniques 1: direct methods.
In: Wilhelm E, Letcher T, editors. Volume properties e
liquids, solutions and vapours. Cambridge: Royal Society of
Chemistry; 2015. p. 73e9.
[46] McLinden MO, Kleinrahm R, Wagner W. Force transmission
errors in magnetic suspension densimeters. Int J
Thermophys 2007;28:429e48. https://doi.org/10.1007/s10765-
007-0176-0.
[47] Richter M, Kleinrahm R. Influence of adsorption and
desorption on accurate density measurements of gas
mixtures. J Chem Thermodyn 2014;74:58e66. https://doi.org/
10.1016/j.jct.2014.03.020.
[48] Bitter F. The magnetic susceptibilities of several organic
gases. Phys Rev 1929;33:389e97. https://doi.org/10.1103/
Phys.Rev.33.389.
[49] Herna
ndez-Go
mez R, Tuma D, Segovia JJ, Chamorro CR.
Experimental determination of (p, r, T) data for binary
mixtures of methane and helium. J Chem Thermodyn
2016;96:1e11. https://doi.org/10.1016/j.jct.2015.12.006.
[50] Herna
ndez-Go
mez R, Tuma D, Pe
rez E, Chamorro CR.
Accurate experimental (p, r, and T) data for the introduction
of hydrogen into the natural gas grid (II): thermodynamic
characterization of the methaneehydrogen binary system
21998 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 2 1 9 8 3 e2 1 9 9 8
from 240 to 350 K and pressures up to 20 MPa. J Chem Eng
Data 2018;63:1613e30. https://doi.org/10.1021/acs.jced.
7b01125.
[51] JCGM 100. Evaluation of measurement data e Guide to the
expression of uncertainty in measurement (GUM). Joint
committees for guides in Metrology. Se vres, France: Bureau
International des Poids et Mesures (BIPM); 2008.
[52] Herna
ndez-Go

mez R, Tuma D, Go
mez-Herna

ndez A,
Chamorro CR. Accurate experimental (p, r, T) data for the
introduction of hydrogen into the natural gas grid:
thermodynamic characterization of the nitrogenehydrogen
binary system from 240 to 350 K and pressures up to 20 MPa. J
Chem Eng Data 2017;62:4310e26. https://doi.org/10.1021/
acs.jced.7b00694.